Federal Reserve Bank of Minneapolis Research Department New Deal Policies and the Persistence of the Great Depression: A General Equilibrium Analysis Harold L. Cole and Lee E. Ohanian∗ Working Paper 597 Revised May 2001 ABSTRACT There are two striking aspects of the recovery from the Great Depression in the United States: the recovery was very weak and real wages in several sectors rose significantly above trend. These data contrast sharply with neoclassical theory, which predicts a strong recovery with low real wages. We evaluate the contribution of New Deal cartelization policies designed to limit competition and increase labor bargaining power to the persistence of the Depression. We develop a model of the bargaining process between labor and firms that occurred with these policies, and embed that model within a multi-sector dynamic general equilibrium model. We find that New Deal cartelization policies are an important factor in accounting for the post-1933 Depression. We also find that the key depressing element of New Deal policies was not collusion per se, but rather the link between paying high wages and collusion. ∗Both, U.C.L.A. and Federal Reserve Bank of Minneapolis. We thank Andrew Atkeson, Tom Holmes, Narayana Kocherlakota, Tom Sargent, Nancy Stokey, seminar participants, and in particular, Ed Prescott for comments. Ohanian thanks the Sloan Foundation and the National Science Foundation for support. The views expressed herein are those of the authors and not necessarily those of the Federal Reserve Bank of Minneapolis or the Federal Reserve System. 1. Introduction There are two striking aspects of the recovery from the Great Depression in the United States. The first is that the recovery was weak. After six years of recovery, real output re- mained 25 percent below trend, and private hours worked were only slightly higher than their 1933 trough level. The second aspect is that real wages in several sectors were significantly above trend, despite the continuation of the depression. The real wage in manufacturing was about 20 percent above trend in 1939, even though manufacturing hours were substantially below trend. These data contrast sharply with neoclassical theory, which predicts a strong recovery from the Great Depression with low real wages, not a weak recovery with high wages. Theory predicts a strong recovery because money, banking, and productivity shocks, which were large and negative before 1933, rebound rapidly after 1933. The pattern of these shocks implies that employment should have returned to trend rapidly, and that wages should have remained below trend throughout the recovery.1 Some economists have suggested that the weak recovery was due to New Deal carteliza- tion policies. These policies permitted industry-wide collusion provided that firms raised wages and agreed to collective bargaining.2 Several empirical studies present evidence these policies increased wages and prices in some sectors during the recovery. However, there is no theoretical model of these policies to understand how they affected the recovery, nor has 1Lucas and Rapping (1972) conclude that rapid money growth should have returned employment and output to normal levels by 1935. Cole and Ohanian (1999) conclude that rapid productivity growth should have returned employment and output to normal levels by 1936, and that wages should have been much lower than observed. Cole and Ohanian (1999) also conclude that other shocks, including financial intermediation shocks, international trade shocks, and public finance shocks can’t reasonably account for the continuation of the Depression. 2See Friedman and Schwartz (1963), Alchian (1970), and Lucas and Rapping (1972). there been any systematic quantitative-theoretic analysis of the impact of these policies on macroeconomic activity. This paper develops a multi-sector dynamic general equilibrium model and uses it to estimate the impact of these policies on employment, output, consumption, and investment between 1934 and 1939. We first develop a model of the intraindustry bargaining process between labor and firms that occurred with these policies. We then embed that bargaining model within the general equilibrium model. We estimate the fraction of the sectors in the economy affected by these policies, and we treat the other sectors of the economy as competitive. We use our model to address two questions about these policies: (1) How distorting are these policies relative to perfect competition, and (2) how much did they contribute to the weak recovery? We address these questions by computing the equilibrium path of the cartel model economy between 1934 and 1939, and comparing it to the equilibrium path of a perfectly competitive version of the model, and also to the data. We find for plausible parameter values that these policies are very distorting - employment and output in the cartel model are about 10 to 15 percent below their counterparts in the perfectly competitive economy. We also find that these policies account for about 60 percent of the weak recovery. The New Deal policies are highly distorting not because of collusion per se, but rather because the policies linked the ability to collude with raising labor bargaining power. This link creates an important insider/outsider friction in our cartel model that raises wages above their competitive levels. In our parameterized model, wages in the cartelized sectors are about 20 percent above trend, as in the data. If wages had not risen, the recovery would have been much faster. 2 The paper is organized as follows. Section 2 presents the data on the recovery for the 1930s. Section 3 discusses the New Deal policies and presents wage and price data from some industries covered by the policies. Section 4 develops competitive and cartel versions of the model economy. In section 5, we choose values for the model’s parameters. Section 6 presents the quantitative analysis. Section 7 describes changes in labor and industrial policies during the 1940s, and the implications of those changes for our model. Section 8 presents a summary and conclusion. 2. The Persistence of the Great Depression This section summarizes data from our earlier paper (Cole and Ohanian (1999), here- after CO) and identifies three puzzles from this period. Table 1 presents the data: real GNP, real consumption of nondurables and services (C), real investment (I), including consumer durables, total factor productivity, (TFP), the real wage in manufacturing, and total private hours worked. All quantities are divided by the adult (16 and over) population, and all vari- ables are measured relative to their trend-adjusted 1929 level. CO describes the data and the detrending procedure in detail.3 CO describe five empirical patterns between 1933 and 1939: (1) GNP and hours worked are significantly below trend. (2) Consumption is flat, remaining about 25 percent below trend over the period. (3) Investment is about 50 percent below trend. (4) Productivity returns to trend by 1936, and remains on trend afterwards. (5) The real wage in manufacturing is significantly above trend. 3Since the theory implies that hours worked are constant along the balanced growth path, we don’t detrend hours per adult, but rather report them relative to their 1929 levels. Unlike CO, we detrend real manufacturing wages by the average growth rate in manufacturing compensation during the postwar period (1.4% per year), rather than the average growth rate of real output per adult (1.9% per year). 3 There are three puzzles about these data: (1) Why was the recovery so weak, given rapid productivity growth? CO show that most U.S. recoveries are rapid, that consumption recovers smoothly to trend, and that investment substantially exceeds its trend level during the recovery phase. (2) Why was the real wage in manufacturing so high during a period of low economic activity?4 (3) Why was labor input so low with such high wages and low consumption? Competitive forces should have led to an equilibrium with higher labor input, a lower real wage, and higher consumption. This coincidence of the high wage, low employment and low consumption suggest some shock seriously distorted the labor market. There are two reasons why the labor market distortion is due to a domestic shock, rather than an international shock. First, CO show most countries had rapid recoveries following the Depression, which is inconsistent with a common international shock. Second, the U.S. had a small trade share during the 1930s. This indicates that the macroeconomic effects of international shocks working through U.S. trade flows would be weak. A successful theory of the recovery should account for the continuation of the De- pression, the increase in the real wage relative to trend, the lack of competition in the labor market, and should be based on a domestic shock. We develop a model driven by a domestic shock that reduces competition in labor and product markets - New Deal labor and industrial policies. We now describe the policies. 4The increase in the real wage during the recovery is not due to imperfectly flexible wages and unanticipated deflation, as has been suggested for the downturn of 1929-1933. Between 1933 and 1939, both nominal wages and the price level increased. Since employment grew after 1933, the wage increase is not easily explained by changes in the average quality of workers. 4 3. New Deal Labor and Industrial Policies Wages and prices in several sectors of the economy rose considerably in mid-1933, and remained high through the decade. Several researchers have argued that the NIRA was responsible for high wages and prices between 1933 and 1935. In this section, we briefly describe NIRA policies and present some data showing wage and price increases during this period. We also summarize post-NIRA labor and industrial policies and argue that labor and industrial policies were responsible for the continuation of high prices and wages between 1935 and 1939.